Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A stroke diagnosis apparatus based on AI (Artificial Intelligence), the apparatus comprising: an image obtainer obtaining a non-contrast CT image related to a brain of at least one patient; a preprocessor pre-processing the non-contrast CT image and determining whether the at least one patient is in a non-hemorrhage state or a hemorrhage state on the basis of the pre-processed image; an image processor normalizing the pre-processed image and dividing and extracting an ROI (Region of Interest) using a preset standard mask template; and a determiner determining whether there is a problem with a cerebral large vessel of the at least one patient using the divided and extracted ROI, wherein the determiner estimates an ASPECT score of the at least one patient using the divided and extracted ROI when there is a problem with the cerebral large vessel of the at least one patient, and determines that the at least one patient is a patient to whom mechanical thrombectomy is applied only when the estimated ASPECT score is a predetermined value or more.
This invention relates to an AI-based stroke diagnosis apparatus designed to assess cerebral large vessel occlusion (LVO) and determine eligibility for mechanical thrombectomy. The apparatus addresses the need for rapid and accurate stroke diagnosis, particularly in identifying patients who may benefit from mechanical thrombectomy, a time-sensitive treatment for ischemic stroke caused by LVO. The apparatus includes an image obtainer that acquires non-contrast CT images of a patient's brain. A preprocessor then processes these images to classify the patient as either in a non-hemorrhage or hemorrhage state. The image processor normalizes the pre-processed images and uses a predefined standard mask template to divide and extract regions of interest (ROIs). A determiner analyzes these ROIs to detect issues with the cerebral large vessels. If a problem is identified, the determiner estimates the patient's ASPECT (Alberta Stroke Program Early CT Score) score, which evaluates the extent of brain damage. Mechanical thrombectomy is recommended only if the ASPECT score meets or exceeds a predetermined threshold, ensuring treatment is administered to patients most likely to benefit. The system automates stroke assessment, reducing diagnostic delays and improving treatment decision-making by leveraging AI-driven image analysis.
2. The stroke diagnosis apparatus according to claim 1 , wherein the preprocessor includes: a noise filter removing noise from the non-contrast CT image; a register performing co-registration spatially aligning images in objects or between a plurality of objects existing in the non-contrast CT image with the noise removed; a skull stripper removing portions that are not a brain structure of the at least one patient from the CT image in which the co-registration has been performed; and a hemorrhage classifier determining whether the at least one patient is in a non-hemorrhage state or a hemorrhage state using the CT image in which skull-stripping has been performed.
This invention relates to a stroke diagnosis apparatus that processes non-contrast CT images to detect hemorrhages in patients. The apparatus addresses the challenge of accurately identifying brain hemorrhages from CT scans, which is critical for timely stroke diagnosis and treatment. The system includes a preprocessor that enhances image quality and extracts relevant brain structures for analysis. First, a noise filter removes artifacts from the non-contrast CT image to improve clarity. Next, a registration module performs co-registration, spatially aligning images within the scan or between multiple scans to ensure consistent anatomical positioning. A skull stripper then removes non-brain structures, isolating the brain for further analysis. Finally, a hemorrhage classifier evaluates the processed image to determine whether the patient is in a non-hemorrhage or hemorrhage state. This automated workflow streamlines stroke diagnosis by reducing manual interpretation errors and accelerating decision-making. The apparatus leverages advanced image processing techniques to enhance diagnostic accuracy in emergency stroke care.
3. The stroke diagnosis apparatus according to claim 2 , wherein when the non-contrast CT image is taken with a gantry inclined, the noise filter performs a tilt correction function that restores an error due to the inclination through re-sampling using gantry tile header information stored together in the original of the non-contrast CT image.
This invention relates to stroke diagnosis using non-contrast CT imaging, addressing the challenge of image distortion caused by gantry inclination during scanning. When a CT scanner's gantry is tilted, the resulting non-contrast CT images may contain geometric errors that affect diagnostic accuracy. The apparatus includes a noise filter that corrects these errors by analyzing gantry tilt header information embedded in the original CT image data. The filter applies a re-sampling technique to restore the correct spatial relationships in the image, ensuring accurate stroke diagnosis. The system may also include a stroke detection module that processes the corrected CT images to identify stroke-related abnormalities, such as hyperdense arteries or hypodense regions, by comparing pixel intensity values against predefined thresholds. Additionally, the apparatus may incorporate a user interface for displaying the corrected images and diagnostic results, along with a data storage system for archiving patient records. The invention improves stroke diagnosis by mitigating gantry-induced distortions and enhancing image clarity for medical professionals.
4. The stroke diagnosis apparatus according to claim 2 , wherein the register spatially aligns images in the objects or between the plurality of objects existing in the non-contrast CT image that are derived by at least one of inclination or a difference in brain shape due to movement of the at least one patient when the non-contrast CT image is taken.
A stroke diagnosis apparatus includes a registration module that spatially aligns images of objects within or between multiple objects in a non-contrast CT image. The alignment corrects for distortions caused by patient movement, such as inclination or variations in brain shape, during the imaging process. This ensures accurate comparison and analysis of anatomical structures across different scans or within a single scan. The apparatus may also include a feature extraction module that identifies and quantifies specific anatomical features, such as blood vessels or tissue regions, from the aligned images. Additionally, a stroke detection module analyzes the extracted features to detect stroke-related abnormalities, such as hemorrhages or ischemic regions, by comparing the features against predefined criteria or reference data. The system may further include a display module that visualizes the detected abnormalities and provides diagnostic information to medical professionals. The alignment process enhances the reliability of stroke diagnosis by minimizing artifacts caused by patient movement, improving the accuracy of feature extraction and subsequent stroke detection.
5. The stroke diagnosis apparatus according to claim 2 , wherein the skull stripper removes portions that are not the brain structure in the CT image on the basis of a skull having a relatively higher Hounsfield unit (HU) value than brain tissues.
A stroke diagnosis apparatus processes CT images to analyze brain structures for stroke detection. The apparatus includes a skull stripper that isolates the brain by removing non-brain structures from the CT image. The skull stripper identifies and removes portions of the image with relatively higher Hounsfield Unit (HU) values, which correspond to the skull, compared to brain tissues. This preprocessing step enhances the accuracy of subsequent stroke diagnosis by ensuring only relevant brain structures are analyzed. The apparatus may also include a brain tissue classifier that categorizes brain regions into different tissue types, such as gray matter, white matter, and cerebrospinal fluid, based on HU values. Additionally, a stroke region detector identifies potential stroke-affected areas by analyzing the classified brain tissues. The stroke diagnosis apparatus may further include a stroke severity evaluator that assesses the severity of detected strokes by measuring the volume and location of the affected regions. The system may also incorporate a user interface for displaying the processed images and diagnostic results to medical professionals. The overall goal is to provide an automated, accurate, and efficient tool for stroke diagnosis using CT imaging.
6. The stroke diagnosis apparatus according to claim 2 , wherein the hemorrhage classifier determines whether the at least one patient is in the non-hemorrhage state or the hemorrhage state using the CT image in which the skullstripping has been performed, on the basis of an A 1 model learning cases related to hemorrhage, and the AI model of the hemorrhage classifier is learned using non-contrast CT data of the at least one patient.
This invention relates to a stroke diagnosis apparatus that analyzes CT images to determine whether a patient is experiencing a hemorrhage. The apparatus includes a hemorrhage classifier that processes CT images where skull-stripping has been performed, meaning the skull has been digitally removed to focus on brain tissue. The classifier uses an AI model trained on cases related to hemorrhage, specifically non-contrast CT data from patients. The AI model evaluates the processed CT images to classify the patient as either in a non-hemorrhage or hemorrhage state. The apparatus leverages machine learning to improve diagnostic accuracy by focusing on relevant brain structures and utilizing non-contrast CT data, which is commonly available in emergency settings. This approach aims to enhance early stroke detection and treatment decisions by automating the identification of hemorrhagic strokes, reducing reliance on manual interpretation and potential human error. The AI model's training on non-contrast CT data ensures compatibility with standard imaging protocols, making the system practical for widespread clinical use.
7. The stroke diagnosis apparatus according to claim 6 , wherein the hemorrhage classifier makes the AI model learn the cases related to the hemorrhage using at least one of Intraparenchymal image, Intraventricular image, Subarachnoid image, Subdural image, and Epidural image in the CT image in which the skull-stripping has been performed.
This invention relates to a stroke diagnosis apparatus that improves the accuracy of hemorrhage detection in computed tomography (CT) images. The apparatus addresses the challenge of accurately identifying different types of hemorrhages in stroke patients, which is critical for timely and effective treatment. The system uses an artificial intelligence (AI) model trained on various hemorrhage-related CT image data, including intraparenchymal, intraventricular, subarachnoid, subdural, and epidural hemorrhage cases. The AI model is trained using CT images that have undergone skull-stripping, a preprocessing step that removes non-brain structures to enhance the focus on brain tissues. By leveraging these specific image types, the apparatus enhances the model's ability to distinguish between different hemorrhage locations and types, improving diagnostic precision. The apparatus may also include additional components such as a stroke classifier for identifying stroke types and a hemorrhage classifier for detecting and classifying hemorrhages. The overall system aims to provide a more reliable and automated diagnostic tool for medical professionals, reducing the risk of misdiagnosis and improving patient outcomes.
8. The stroke diagnosis apparatus according to claim 6 , wherein the hemorrhage classifier constructs the AI model by configuring a convolutional neural network (CNN) to extract feature maps for each data slice of the non-contrast CT data of the at least one patient, and applying and combining the extracted plurality of feature maps sequentially with a Long Short-term Memory (LSTM) neural network.
This invention relates to a stroke diagnosis apparatus that analyzes non-contrast CT scans to detect hemorrhagic strokes. The apparatus includes a hemorrhage classifier that constructs an AI model using a convolutional neural network (CNN) to process each slice of the CT data. The CNN extracts feature maps from each slice, capturing spatial patterns in the medical images. These feature maps are then sequentially fed into a Long Short-term Memory (LSTM) neural network, which analyzes the temporal relationships between slices to identify hemorrhage indicators. The combination of CNN and LSTM allows the model to leverage both spatial and sequential information from the CT scans, improving accuracy in stroke detection. The apparatus may also include a preprocessing module to enhance image quality and a user interface for displaying diagnostic results. The system is designed to assist clinicians in rapidly and accurately diagnosing hemorrhagic strokes, reducing the time between imaging and treatment.
9. The stroke diagnosis apparatus according to claim 2 , wherein the hemorrhage classifier determines that the at least one patient is in the hemorrhage state when a specific factor exists in over predetermined regions in the CT image in which the skull-stripping has been performed regardless of the hemorrhage classification.
This invention relates to a stroke diagnosis apparatus designed to improve the accuracy of detecting hemorrhages in computed tomography (CT) images. The apparatus addresses the challenge of reliably identifying hemorrhages in stroke patients, where conventional methods may fail to account for certain factors that indicate bleeding regardless of predefined hemorrhage classifications. The apparatus includes a hemorrhage classifier that analyzes CT images after skull-stripping, a preprocessing step that removes non-brain structures to focus on brain tissue. The classifier evaluates the presence of specific factors, such as abnormal blood density or contrast patterns, across multiple regions of the brain. If these factors are detected in more than a predetermined number of regions, the apparatus concludes that the patient is in a hemorrhage state, even if the factors do not match traditional hemorrhage classifications. This approach enhances detection sensitivity by considering broader indicators of bleeding, reducing false negatives in stroke diagnosis. The apparatus may also include a stroke classifier that distinguishes between ischemic and hemorrhagic strokes based on image features, and a stroke state classifier that determines the severity of the stroke. The hemorrhage classifier operates independently, ensuring that critical bleeding is identified even if other stroke-related factors are ambiguous. This multi-layered analysis improves diagnostic accuracy and supports timely medical intervention.
10. The stroke diagnosis apparatus according to claim 1 , wherein the image processor includes: a template setter setting the standard mask template in advance; an image normalizer normalizing the pre-processed image; and an image processor dividing and extracting the ROI by applying the preset standard mask template to the normalized image; wherein the image processor operates only when it is determined that the at least one patient is in the non-hemorrhage state.
This invention relates to a stroke diagnosis apparatus designed to analyze medical images for stroke detection, specifically focusing on identifying non-hemorrhagic stroke conditions. The apparatus includes an image processor that enhances diagnostic accuracy by selectively operating only when a patient is confirmed to be in a non-hemorrhage state, avoiding false positives from hemorrhage-related artifacts. The image processor comprises three key components: a template setter that predefines a standard mask template, an image normalizer that standardizes pre-processed medical images, and an image processor that applies the template to the normalized image to extract regions of interest (ROIs). The template-based extraction ensures consistent and precise analysis of relevant brain regions, improving the reliability of stroke detection. The apparatus is particularly useful in clinical settings where rapid and accurate differentiation between hemorrhagic and non-hemorrhagic strokes is critical for treatment decisions. By automating the ROI extraction process and conditioning its operation on the absence of hemorrhage, the invention streamlines diagnostic workflows and reduces the risk of misdiagnosis. The system leverages pre-processing techniques to prepare images for analysis, ensuring optimal conditions for template application and ROI extraction. This selective activation mechanism enhances diagnostic efficiency and accuracy, making it a valuable tool for stroke diagnosis in medical imaging.
11. The stroke diagnosis apparatus according to claim 10 , wherein the template setter collects a plurality of medical images obtained from a plurality of normal people and patients with brain diseases, creates a 3D normalization image on the basis of the collected images, creates a 2D normalization image by slicing the 3D normalization image based on one axis of an X-axis, a Y-axis, and a Z-axis on the basis of 3D voxels that are predetermined units, and sets the standard mask template in advance on the basis of the ROIs divided from the created 2D normalization image.
This invention relates to a stroke diagnosis apparatus that improves the accuracy of stroke detection by using standardized brain image templates. The apparatus addresses the challenge of variability in brain imaging data, which can lead to misdiagnosis or delayed treatment. The system collects medical images from both healthy individuals and patients with brain diseases, such as stroke, to create a 3D normalization image. This 3D image is then sliced along one of the X, Y, or Z axes to generate a 2D normalization image, using predefined 3D voxel units. The apparatus divides the 2D image into regions of interest (ROIs) and sets a standard mask template based on these ROIs. This template serves as a reference for comparing new patient scans, allowing for more consistent and reliable stroke diagnosis. The use of normalized images and predefined ROIs helps reduce errors caused by anatomical differences between individuals, improving diagnostic accuracy. The apparatus may also include additional features, such as a display for visualizing the comparison between patient scans and the template, and a processor for analyzing the images. The overall system enhances stroke detection by standardizing brain imaging data and providing a structured framework for analysis.
12. The stroke diagnosis apparatus according to claim 10 , wherein the image normalizer performs the normalization by changing a Hounsfield unit (HU) value of the pre-processed image.
This invention relates to stroke diagnosis using medical imaging, specifically addressing the challenge of accurately analyzing brain images to detect stroke-related abnormalities. The apparatus includes an image normalizer that standardizes pre-processed medical images, such as CT scans, by adjusting their Hounsfield Unit (HU) values. HU values represent tissue density in CT imaging, but variations in scanning parameters or patient conditions can lead to inconsistent readings, complicating stroke diagnosis. The normalization process ensures consistent HU values across different scans, improving the reliability of subsequent stroke detection algorithms. The apparatus also includes a stroke detector that analyzes the normalized images to identify stroke indicators, such as regions of abnormal density or blood flow disruption. The system may further incorporate a pre-processor to enhance image quality before normalization, addressing issues like noise or artifacts. By standardizing HU values, the apparatus enhances the accuracy and consistency of stroke diagnosis, enabling earlier and more reliable detection of ischemic or hemorrhagic strokes. The invention is particularly useful in clinical settings where rapid and precise stroke assessment is critical for patient outcomes.
13. The stroke diagnosis apparatus according to claim 1 , wherein the determiner includes: an ischemia classifier determining whether there is ischemia in the brain of the at least one patient using the divided and extracted ROI; a large vessel occlusion determiner determining whether there is a problem with a cerebral large vessel of the at least one patient when there is the ischemia; an ASPECTS determiner estimating an ASPECT score of the at least one patient using the divided and extracted ROI when there is a problem with the cerebral large vessel; and a thrombectomy determiner determining that the at least one patient is a patient to whom the mechanical thrombectomy is applied only when the estimated ASPECT score is a predetermined value or more.
This invention relates to a stroke diagnosis apparatus designed to assess and diagnose ischemic stroke in patients, particularly focusing on identifying large vessel occlusion and determining eligibility for mechanical thrombectomy. The apparatus processes medical imaging data, such as CT or MRI scans, to extract and analyze regions of interest (ROIs) in the brain. The system includes an ischemia classifier that detects the presence of ischemia in the brain. If ischemia is detected, a large vessel occlusion determiner evaluates whether a cerebral large vessel is affected. If a large vessel problem is identified, an ASPECTS (Alberta Stroke Program Early CT Score) determiner calculates the patient's ASPECT score, which quantifies the extent of brain damage. Finally, a thrombectomy determiner assesses whether the patient is suitable for mechanical thrombectomy based on the ASPECT score, recommending the procedure only if the score meets or exceeds a predefined threshold. This structured approach ensures rapid and accurate stroke diagnosis, aiding in timely treatment decisions.
14. The stroke diagnosis apparatus according to claim 13 , wherein the ischemia classifier determines whether there is the ischemia on the basis of whether there is an affected part in the divided and extracted ROI and an AI model learning shape classification, and the determination of whether there is the affected part in the divided and extracted ROI and the shape classification is performed using Old infarct ( 0 I), Recent infarct (RI), Frank hypodensity (FH), and Early ischemic sign (EIS).
15. The stroke diagnosis apparatus according to claim 14 , wherein the ischemia classifier determines that there is the ischemia when there is at least one of the Old infarct ( 0 I), Recent infarct (RI), Frank hypodensity (FH), and Early ischemic sign (EIS) in any one region of the divided and extracted ROIs.
This invention relates to a stroke diagnosis apparatus designed to detect ischemic stroke by analyzing brain imaging data. The apparatus addresses the challenge of early and accurate stroke diagnosis, which is critical for timely medical intervention. The system processes brain imaging data, such as CT scans, to identify specific indicators of ischemia in regions of interest (ROIs) within the brain. The apparatus divides the brain into multiple ROIs and evaluates each for the presence of at least one of four key ischemic markers: Old Infarct (OI), Recent Infarct (RI), Frank Hypodensity (FH), or Early Ischemic Sign (EIS). If any of these markers are detected in a single ROI, the apparatus classifies the region as ischemic, enabling rapid diagnosis. The system enhances diagnostic accuracy by focusing on localized ischemic signs rather than relying on broader imaging assessments. This approach improves early detection and treatment planning for stroke patients, reducing the risk of long-term neurological damage. The apparatus integrates advanced image processing and machine learning techniques to automate the detection of ischemic indicators, making it a valuable tool for medical professionals in emergency and diagnostic settings.
16. The stroke diagnosis apparatus according to claim 13 , wherein the problem with the cerebral large vessel is large vessel occlusion, and the large vessel occlusion determiner determines whether there is a possibility of the large vessel occlusion on the basis of whether a dense MCA sign has been detected at an infra-ganglionic level in relation to the divided and extracted ROI.
17. The stroke diagnosis apparatus according to claim 16 , wherein when the dense MCA sign is not detected, the large vessel occlusion determiner detects a frequency of each Hounsfield unit (HU) value in both left and right hemispheres of sequential slice images at the infra-ganglionic level in relation to the divided and extracted ROI, and determines that there is the large vessel occlusion when the frequency of the detected HU values of at least one of the both hemispheres is a predetermined reference value or more or when a frequency difference of the detected HU values of the both hemispheres is a predetermined difference value or more.
This invention relates to a stroke diagnosis apparatus designed to detect large vessel occlusions in the brain, particularly when a dense middle cerebral artery (MCA) sign is not present. The apparatus analyzes sequential slice images of the brain at the infra-ganglionic level, focusing on regions of interest (ROIs) that have been divided and extracted from the images. When the dense MCA sign is absent, the apparatus evaluates the frequency distribution of Hounsfield unit (HU) values in both the left and right hemispheres. It determines the presence of a large vessel occlusion if the frequency of HU values in at least one hemisphere exceeds a predetermined reference value or if the frequency difference between the two hemispheres exceeds a predetermined threshold. This method enhances stroke diagnosis by providing an alternative detection approach when traditional indicators like the dense MCA sign are not evident. The apparatus improves diagnostic accuracy by leveraging quantitative analysis of HU value distributions in brain imaging.
18. The stroke diagnosis apparatus according to claim 13 , wherein the ASPECTS determiner estimates the ASPECT score on the basis of whether there is the affected part in the divided and extracted ROI and an AI model learning shape classification, and the determination of whether there is the affected part in the divided and extracted ROI and the shape classification is performed using Old infarct (OI), Recent infarct (RI), Frank hypodensity (FH), and Early ischemic sign (EIS).
This invention relates to a stroke diagnosis apparatus that analyzes brain imaging data to assess stroke severity using the Alberta Stroke Program Early CT Score (ASPECTS). The apparatus addresses the challenge of accurately and efficiently evaluating early ischemic changes in computed tomography (CT) scans, which is critical for timely stroke diagnosis and treatment. The system includes an ASPECTS determiner that estimates the ASPECT score by analyzing regions of interest (ROIs) in the brain, which are divided and extracted from the imaging data. The determination process involves identifying affected brain regions and classifying their shapes using an artificial intelligence (AI) model trained for shape classification. The AI model distinguishes between different types of stroke-related abnormalities, including Old infarct (OI), Recent infarct (RI), Frank hypodensity (FH), and Early ischemic sign (EIS). By integrating these classifications, the apparatus provides a quantitative assessment of stroke severity, aiding clinicians in rapid decision-making. The system enhances diagnostic accuracy by automating the analysis of subtle imaging features that may be difficult to detect manually, thereby improving patient outcomes.
19. The stroke diagnosis apparatus according to claim 18 , wherein when at least one of the Old infarct ( 0 I), Recent infarct (RI), Frank hypodensity (FH), and Early ischemic sign (EIS) is detected, the ASPECTS determiner admits it as the affected part and reflects the value admitted as the affected part to the estimation of the ASPECT score.
This invention relates to stroke diagnosis, specifically improving the accuracy of the Alberta Stroke Program Early CT Score (ASPECTS) assessment by automatically detecting and incorporating specific brain abnormalities into the scoring system. The ASPECTS score is used to evaluate early ischemic stroke severity on CT scans, but traditional methods rely on manual interpretation, which can be inconsistent. The apparatus addresses this by automatically identifying and classifying stroke-related features in CT images, including old infarcts (OI), recent infarcts (RI), frank hypodensity (FH), and early ischemic signs (EIS). When any of these features are detected in a brain region, the system designates that region as affected and adjusts the ASPECT score accordingly. This automated approach reduces human error and provides a more objective assessment of stroke severity, aiding in faster and more accurate clinical decisions. The system integrates these detections into the ASPECTS calculation, ensuring that all relevant abnormalities contribute to the final score, which is critical for determining treatment options and prognosis. The invention enhances the reliability of stroke diagnosis by leveraging automated image analysis to standardize the evaluation process.
20. The stroke diagnosis apparatus according to claim 18 , wherein when the Frank hypodensity (FH) and Early ischemic sign (EIS) of the Old infarct (OI), Recent infarct (RI), Frank hypodensity (FH), and Early ischemic sign (EIS) are detected, the ASPECTS determiner admits them as the affected part and reflects the value admitted as the affected part to the estimation of the ASPECT score.
This invention relates to a stroke diagnosis apparatus designed to improve the accuracy of stroke assessment by analyzing brain imaging data. The apparatus addresses the challenge of reliably detecting and quantifying stroke-affected regions in computed tomography (CT) scans, particularly for ischemic strokes, where early detection is critical for treatment decisions. The apparatus includes an ASPECTS (Alberta Stroke Program Early CT Score) determiner that evaluates specific brain regions to assess stroke severity. The invention enhances this process by detecting and classifying different types of brain abnormalities, including Frank hypodensity (FH), Early ischemic sign (EIS), Old infarct (OI), and Recent infarct (RI). When these abnormalities are identified, the apparatus designates them as affected regions and incorporates their presence into the ASPECT score calculation. This ensures that all relevant stroke indicators are considered, improving diagnostic precision. The apparatus may also include a brain region detector to identify the 10 regions of the brain typically assessed in the ASPECT score, a stroke type classifier to distinguish between different stroke-related abnormalities, and a display to present the ASPECT score and affected regions. The system automates the analysis of CT scans, reducing human error and providing faster, more consistent stroke evaluations. This technology supports clinicians in making timely treatment decisions, particularly in acute stroke care.
21. The stroke diagnosis apparatus according to claim 18 , wherein the divided and extracted ROI includes an MCA (Middle cerebral artery), ACA (Anterior cerebral artery), PCA (Posterior cerebral artery), and ICA (Internal carotid artery) regions, and when the Frank hypodensity (FH) and the Early ischemic sign (EIS) of the Old infarct ( 0 I), Recent infarct (RI), Frank hypodensity (FH), and Early ischemic sign (EIS) are detected in the MCA, ACA, PCA, and ICA regions, and the ASPECTS determiner admits them as the affected part and reflects the values admitted as the affected parts to the estimation of the ASPECT score.
This invention relates to a stroke diagnosis apparatus designed to improve the accuracy of stroke assessment by analyzing specific regions of the brain's vasculature. The apparatus focuses on detecting and evaluating ischemic stroke indicators in key cerebral arteries, including the Middle Cerebral Artery (MCA), Anterior Cerebral Artery (ACA), Posterior Cerebral Artery (PCA), and Internal Carotid Artery (ICA). The system extracts regions of interest (ROIs) from medical imaging data, such as CT scans, and identifies stroke-related abnormalities like Frank hypodensity (FH) and Early Ischemic Signs (EIS) in both old and recent infarcts. These detected abnormalities are mapped to the corresponding arterial regions. The apparatus then applies the Alberta Stroke Program Early CT Score (ASPECTS) methodology, assigning scores based on the presence of these abnormalities in the defined regions. The ASPECT score, which quantifies the extent of affected brain tissue, is adjusted according to the detected abnormalities, providing a more precise stroke diagnosis. This approach enhances the reliability of stroke assessment by systematically analyzing multiple arterial regions and their associated ischemic signs, aiding in timely and accurate treatment decisions.
22. The stroke diagnosis apparatus according to claim 13 , wherein, when the estimated ASPECT score is a predetermined value or more, the thrombectomy determiner determines that the at least one patient is a patient to whom the mechanical thrombectomy is applied, only when a reference determined through information obtained by the ischemia classifier and information obtained by the large vessel occlusion determiner is a predetermined reference or more.
This invention relates to a stroke diagnosis apparatus designed to improve the accuracy of determining whether a patient is suitable for mechanical thrombectomy, a procedure used to remove blood clots in ischemic stroke cases. The apparatus addresses the challenge of ensuring that thrombectomy is performed only on patients who will benefit, as the procedure carries risks and is not universally effective. The apparatus includes an ischemia classifier that analyzes medical data to determine if a patient has ischemic stroke, a large vessel occlusion determiner that identifies whether a large blood vessel is blocked, and an ASPECT score estimator that evaluates the extent of brain tissue affected by the stroke. The key innovation is a thrombectomy determiner that uses these components to decide if a patient should undergo thrombectomy. Specifically, if the estimated ASPECT score meets or exceeds a predetermined threshold, the apparatus further checks a reference value derived from both the ischemia classifier and large vessel occlusion determiner. Only if this reference value also meets or exceeds a second predetermined threshold does the apparatus conclude that the patient is a suitable candidate for thrombectomy. This dual-threshold approach aims to enhance decision-making by combining multiple diagnostic factors, reducing unnecessary procedures and improving patient outcomes.
23. The stroke diagnosis apparatus according to claim 13 , wherein the thrombectomy determiner determines whether the at least one patient is a patient to whom the mechanical thrombectomy is applied, on the basis of an AI model learning a volume value detected due to a problem with the cerebral large vessel.
This invention relates to a stroke diagnosis apparatus designed to assess whether a patient is suitable for mechanical thrombectomy, a procedure used to remove blood clots from cerebral large vessels. The apparatus addresses the challenge of accurately identifying patients who would benefit from this intervention, improving treatment outcomes by ensuring appropriate patient selection. The apparatus includes a thrombectomy determiner that evaluates patient data using an artificial intelligence (AI) model. The AI model is trained on volume values associated with cerebral large vessel issues, such as clot size or vessel occlusion severity. By analyzing these volume values, the model determines whether a patient is a candidate for mechanical thrombectomy. This decision-making process helps clinicians prioritize treatment options and optimize stroke care. The system integrates with broader diagnostic tools, such as imaging devices, to collect relevant patient data. The AI model processes this data to provide a binary or probabilistic output indicating thrombectomy suitability. This approach enhances precision in stroke diagnosis and treatment planning, reducing reliance on subjective assessments and improving patient outcomes. The invention aims to streamline decision-making in acute stroke management, ensuring timely and effective interventions.
24. The stroke diagnosis apparatus according to claim 13 , wherein the thrombectomy determiner calculates a tissue clock using an absolute time, the volume of Early ischemic sign (EIS), and the volume of Frank hypodensity (FH) until the non-contrast CT imaging after a problem with the cerebral large vessel is generated, and determines whether the at least one patient is a patient to whom the mechanical thrombectomy is applied, on the basis of the calculated tissue clock.
This invention relates to stroke diagnosis, specifically improving the determination of whether a patient is suitable for mechanical thrombectomy, a procedure to remove blood clots from cerebral large vessels. The challenge is accurately assessing the timing and severity of ischemic damage to decide if thrombectomy is viable within the critical treatment window. The apparatus calculates a "tissue clock" using three key parameters: the absolute time from symptom onset to non-contrast CT imaging, the volume of Early Ischemic Signs (EIS), and the volume of Frank Hypodensity (FH). EIS refers to subtle early signs of ischemia, while FH indicates more advanced tissue damage. By integrating these factors, the system quantifies the extent of irreversible brain injury and estimates the remaining viable tissue at risk. The thrombectomy determiner then evaluates whether the patient meets criteria for mechanical thrombectomy based on this tissue clock, ensuring treatment is applied only when likely to be effective. This approach enhances decision-making by providing a data-driven assessment of ischemic progression, reducing reliance on subjective clinical judgment and improving patient outcomes by optimizing thrombectomy eligibility.
25. The stroke diagnosis apparatus according to claim 1 , further comprising a communication unit transmitting information about the at least one patient to a pre-designated outside when the at least one patient is a patient to whom the mechanical thrombectomy is applied.
This invention relates to a stroke diagnosis apparatus designed to improve the efficiency and accuracy of stroke diagnosis and treatment, particularly for patients undergoing mechanical thrombectomy. The apparatus includes a diagnostic module that analyzes patient data to determine the likelihood of a stroke and identifies patients who may require mechanical thrombectomy, a treatment involving the removal of blood clots from blood vessels. The apparatus further includes a communication unit that automatically transmits patient information to a pre-designated external entity, such as a hospital or medical database, when the patient is confirmed to be undergoing mechanical thrombectomy. This ensures timely sharing of critical patient data with healthcare providers, facilitating faster decision-making and coordination in stroke treatment. The system may also include a display unit to present diagnostic results and treatment recommendations to medical professionals. The communication unit may use wired or wireless transmission methods to relay data securely. The apparatus is intended to streamline stroke diagnosis and treatment workflows, reducing delays and improving patient outcomes by ensuring that relevant medical information is promptly shared with the appropriate parties.
26. The stroke diagnosis apparatus according to claim 25 , the information that is transmitted to the outside includes elapsed time information, non-contrast CT image information, determination result information, tissue clock information, and the estimated ASPECT score information that are related to the at least one patient.
This invention relates to a stroke diagnosis apparatus designed to improve the accuracy and efficiency of stroke assessment. The apparatus addresses the critical need for rapid and precise stroke diagnosis, particularly in determining the extent of brain tissue damage and viability for treatment decisions. The system captures and processes non-contrast CT image data from at least one patient, analyzing the images to generate diagnostic information. Key outputs include elapsed time information (tracking the duration since symptom onset), non-contrast CT image data, determination results (likely indicating stroke presence or severity), tissue clock information (assessing the viability of brain tissue over time), and estimated ASPECT score information (a standardized scoring system for evaluating ischemic stroke damage). The apparatus transmits this comprehensive diagnostic data to external systems or healthcare providers, enabling faster decision-making and treatment planning. The integration of multiple data points, including time-sensitive metrics and standardized scoring, enhances the reliability of stroke assessments, supporting timely interventions such as thrombolytic therapy or mechanical thrombectomy. The system is particularly valuable in emergency settings where rapid diagnosis and treatment initiation are crucial for patient outcomes.
Unknown
January 26, 2021
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.